[FRIAM] Swarm again

[Marcus Daniels]

Coupling ABMs to LLMs have another benefit which is to provide explainable models. Machine learning models are often criticized for being predictive but not explanatory. However, a LLM agent that creates minimal idiomatic computer programs can work with an agent that compiles and runs those programs and can work with an agent that fits real-world data against those simulations. It’s not so different from genetic programming workflows, but the programming productivity is higher. 

From: Friam <friam-bounces at redfish.com> on behalf of Pieter Steenekamp <pieters at randcontrols.co.za>
Date: Saturday, November 9, 2024 at 8:27 AM
To: The Friday Morning Applied Complexity Coffee Group <friam at redfish.com>
Subject: Re: [FRIAM] Swarm again 

Yes, interactions between not only LLMs but also other configurations can indeed lead to emergence. In fact, using an Agent Based Model (ABM) (like in swarm.org <http://swarm.org>) to generate complex data and then training a deep learning model on that data holds a lot of promise. A good example is how DeepMind approached predicting plasma behavior to contain nuclear fusion reactions. Traditional models couldn’t accurately predict plasma behavior, but by first using a swarm.org-type model to generate emergence to genrate training data, they were able to build an AI deep learning model that was used to contain the plasma.
I would not be surprised if this approach is developed sometime in future to manage complexity more efficiently. 


On Sat, 9 Nov 2024 at 17:35, Pietro Terna <pietro.terna at unito.it <mailto:pietro.terna at unito.it>> wrote: 

Sure, but it is not impossible to have the OpenAI agents that interact and generate emergent effects, with the possibility of having them interacting in a deep way using natural language.

Best, Pietro (a rather old guy, formerly Swarm programmer for ABM)


Il 09/11/24 16:19, Pieter Steenekamp ha scritto: 

The two "swarms" are quite different.

Swarm.org aims to create complex behaviors from simple interactions between simple agents. 

On the other hand, OpenAI's swarm involves many different AI agents interacting with each other. However, it doesn't specifically focus on emergence of complexity, which is the main idea behind swarm.org <_blank>. 


On Sat, 9 Nov 2024 at 16:49, Pietro Terna <pietro.terna at unito.it <_blank>> wrote: 

For people who created http://www.swarm.org <_blank> or used it, have a look at 
https://github.com/openai/swarm/blob/main/README.md <_blank>

Two completely different worlds or a unique one?

Best, Pietro


-- Democrito, che ’l mondo a caso pone (Inferno, IV, 136) "It is the hallmark of any deep truth that its negation is also a deep truth." Niels Bohr. A https://terna.to.it/breviArticoli.html <_blank> riporto dei miei brevi articoli su temi di attualità. Asterischi da Verso Itaca a https://verso-itaca.it/asterischi-da-verso-itaca/ <_blank> Home page: https://terna.to.it <_blank> Twitter: https://twitter.com/@pietroterna <_blank> Mastodon: https://mastodon.uno/@PietroTerna <_blank> 
.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ... --- -- . / .- .-. . / ..- ... . ..-. ..- .-..
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom https://bit.ly/virtualfriam <_blank>
to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com <_blank>
FRIAM-COMIC http://friam-comic.blogspot.com/ <_blank>
archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/ <_blank>
1/2003 thru 6/2021 http://friam.383.s1.nabble.com/ <_blank> 




.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ... --- -- . / .- .-. . / ..- ... . ..-. ..- .-.. FRIAM Applied Complexity Group listserv Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom https://bit.ly/virtualfriam <_blank> to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com <_blank> FRIAM-COMIC http://friam-comic.blogspot.com/ <_blank> archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/ <_blank> 1/2003 thru 6/2021 http://friam.383.s1.nabble.com/ <_blank> 


-- Democrito, che ’l mondo a caso pone (Inferno, IV, 136) "It is the hallmark of any deep truth that its negation is also a deep truth." Niels Bohr. A https://terna.to.it/breviArticoli.html <_blank> riporto dei miei brevi articoli su temi di attualità. Asterischi da Verso Itaca a https://verso-itaca.it/asterischi-da-verso-itaca/ <_blank> Home page: https://terna.to.it <_blank> Twitter: https://twitter.com/@pietroterna <_blank> Mastodon: https://mastodon.uno/@PietroTerna <_blank> 
.- .-.. .-.. / ..-. --- --- - . .-. ... / .- .-. . / .-- .-. --- -. --. / ... --- -- . / .- .-. . / ..- ... . ..-. ..- .-..
FRIAM Applied Complexity Group listserv
Fridays 9a-12p Friday St. Johns Cafe / Thursdays 9a-12p Zoom https://bit.ly/virtualfriam <_blank>
to (un)subscribe http://redfish.com/mailman/listinfo/friam_redfish.com <_blank>
FRIAM-COMIC http://friam-comic.blogspot.com/ <_blank>
archives: 5/2017 thru present https://redfish.com/pipermail/friam_redfish.com/ <_blank>
1/2003 thru 6/2021 http://friam.383.s1.nabble.com/ <_blank> 


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20241109/66c128a1/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 8422 bytes
Desc: not available
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20241109/66c128a1/attachment.bin>